Dell, EMC, Dell Technologies, Cisco,

Showing posts with label NVMe. Show all posts
Showing posts with label NVMe. Show all posts

Wednesday, July 11, 2018

IBM's New FlashSystem 9100 All-NVMe Flash Storage Comes Loaded

@IBM on Tuesday unveiled its first storage system featuring embedded high-performance #NVMe storage capacity along with the full complement of IBM's storage software and new storage "blueprints" to help partners build multi-cloud solutions. The new FlashSystem 9100 supports both IBM's proprietary high-performance FlashCore flash storage modules as well as industry-standard SSDs, and comes with Storage Insights, IBM's artificial intelligence-powered analytics software, said Eric Herzog, chief marketing officer and vice president of worldwide storage channels for IBM storage and software-defined infrastructure. "This is not just an NVMe array with faster speed," Herzog told @CRN. "All our competitors are doing this. This ships with tons of software."

https://www.crn.com/news/storage/300106236/ibms-new-flashsystem-9100-all-nvme-flash-storage-comes-loaded.htm

Monday, April 30, 2018

NVMe™ and NVMe-oF™ Take Center Stage at Flash Memory Summit

SANTA CLARA, Calif., April 30, 2018 (GLOBE NEWSWIRE) -- The rapidly emerging #NVMe and #NVMeoF standards for high speed data storage will play a lead role in the 2018 Flash Memory Summit ( #FMS ), scheduled for August 7-9, 2018 at the Santa Clara Convention Center.   According to G2M Research, the NVMe market is expected to reach $60 billion by 2021 as designers make it their preferred solution for high-speed storage. NVMe offers data transfer rates up to 10 times greater than SATA, runs on the popular PCIe interface and offers a rapidly growing ecosystem of products and technology.  The NVMe-oF standard offers similar advantages for networked storage, allowing applications to use flash anywhere in an enterprise network or computer system.  “NVMe should be designers’ first choice among storage interfaces,” said Chuck Sobey, Flash Memory Summit Conference Chair.  “It offers much higher throughput than disk interfaces while still using an interface familiar to all data centers.  NVMe and NVMe-oF meet the needs of enterprise, computer, high-performance computing, embedded, mil/aero, mobile, and many other applications as well.” FMS 2018 will present a range of in-depth technical sessions on NVMe and NVMe-oF, including a series organized by the standards Group NVM Express Inc. as well as pre-conference seminars, invited talks, exhibits and market trends. The Flash Memory Summit will also feature 3D NAND flash and 3D XPoint, emerging nonvolatile memory technologies including applications for the latest MRAM developments, persistent memory and NVDIMM, controllers, flash storage networking, and the software that makes it all work together.  The FMS conference offers an upgraded exhibit hall, including a large demonstration area and more than 100 exhibitors. The conference is still accepting exhibitor applications for booth space and proposals for panel sessions and speaking slots.

https://globenewswire.com/news-release/2018/04/30/1489650/0/en/NVMe-and-NVMe-oF-Take-Center-Stage-at-Flash-Memory-Summit.html

Tuesday, March 27, 2018

How to transition to NVMe SSDs from SAS/SATA

Getting ahead of the SSD interface performance curve requires foresight and careful planning. As NVMe technology evolves, new adopters are looking for ways to ease the changeover. Nonvolatile memory express has emerged as the open logical device interface specification that future generations
of storage systems will be built upon, promising significant performance acceleration at economics that cannot be equaled by earlier generation SAS/SATA systems. In fact, when compared with NVMe SSDs, SAS and SATA have become legacy design interfaces. And IT experts are taking notice.
A growing number of enterprises are beginning to recognize that, if they truly want to benefit from SSD performance, they need to move toward NVMe technology with a direct attachment to the Peripheral Component Interconnect Express (PCIe) bus.
"This ... gets you as close as possible to the CPU," explained Ian McClarty, president of PhoenixNAP Global IT Services. The approach, he added, dramatically reduces latency and increases speeds through multiple PCIe lanes, thereby relieving the bottlenecks faced by SAS/SATA interface users.
Making the move
The transition to NVMe SSD technology will likely take place gradually over the next several years. Luckily for current SAS/SATA users, they don't have to face the immediate prospect of ripping and replacing their existing storage technology or attempting a forklift upgrade. "Instead, [new adopters] should look to implement future-proof architectures that will allow them to scale with new innovations in flash technology," said Jeff Baxter, chief evangelist of ONTAP at storage and data management software provider NetApp. "Some existing NVMe-ready solutions will allow customers to nondisruptively add new NVMe storage arrays directly into existing scale-out architectures and expand or move workloads as appropriate from current HDD or SAS/SATA-attached SSD arrays to NVMe SSD-based arrays over time," he added.
Before launching an NVMe transition, IT leaders should first determine whether their incumbent storage array is NVMe-capable. "In other words, is the array capable of seamlessly adding #NVMe SSDs without requiring a major reconfiguration or data migration?" asked @NarayanVenkat, @WesternDigital 's vice president of #datacenter systems for @Tegile #flash storage arrays. "If so, then simply replacing a few of the SAS/SATA SSDs with NVMe provides the benefits of instant performance acceleration without any disruption to applications."
Unfortunately, Venkat added, most incumbent storage systems aren't capable of seamless NVMe SSD support, since the products weren't designed from the outset with NVMe SSD compatibility in mind.
Navigating compatibility pitfalls
If an incumbent storage array isn't fully NVMe-compatible, as is typically the case, the only practical solution is to upgrade to an NVMe-based storage system that has been designed from the outset to take advantage of the speed and performance of NVMe SSDs. With this approach, managers can easily migrate applications and data to their new NVMe-based storage systems. "Migration is simple for virtualized environments, like VMware, as administrators can move virtual machines, along with applications and data, with a few clicks," Venkat observed.

Monday, March 12, 2018

Eideticom, IBM, Rackspace and Xilinx Demonstrate World's First PCIe Gen4 NVM Express Production Ready System

CALGARY, March 12, 2018 /PRNewswire/ - @Eideticom, @IBM, @Rackspace and @Xilinx are pleased to demonstrate the Eideticom NoLoad™ NVM Express ( #NVMe) Computational Storage acceleration platform running the NVMe protocol at PCIe Gen4 speeds in a Rackspace BarrelEye G2 POWER9-equipped system, doubling performance compared to other systems in production today. This announcement enables Rackspace, IBM and other Open Compute and OpenPOWER system vendors to avail of the next generation of storage and compute acceleration performance leveraging the industry-exclusive PCIe Gen4 capabilities of IBM's POWER9 processor, Rackspace's server, Xilinx's FPGA technology and Eideticom's NVMe-based accelerator products. Eideticom, IBM, Rackspace and Xilinx Demonstrate World's First PCIe Gen4 NVM Express Production Ready System (CNW Group/Eidetic Communications Inc) NVMe is the most popular protocol for Solid State Drives (SSDs) in Enterprise, Data-Center and High-Performance Computing markets. It has a broad eco-system and excellent software support in all major operating systems including Linux. NVMe uses PCIe to connect the CPU to the SSDs and IBM's POWER9 is the first production CPU with PCIe Gen4 IO, nearly doubling the data bandwidth of PCIe Gen3, to which x86 remains committed. Eideticom deployed their NVMe-based accelerator, NoLoad™ product on top of Xilinx's FPGA technology on a production ready FPGA acceleration card inside a production ready OpenPOWER server from Rackspace. "We built the POWER9 CPU from the ground-up with the most data-intensive workloads in mind," said IBM VP and Fellow Brad McCredie. "PCIe Gen4 devices are critical components to driving performance across the system stack, and with POWER9 their full potential can be realized." "We are excited to incorporate Eideticom's storage acceleration and PCIe Gen4 technology in our Barreleye G2 server," said Adi Gangidi, system design engineer at Rackspace. "Accelerator IP enablers like Eideticom are helping drive the widespread data center adoption of a new and unmatched class of IO." "Xilinx UltraScale+ All Programmable devices offer unique capabilities to enable the highest performance for demanding Data Center storage and compute acceleration workloads," said Manish Muthal, Vice President of Data Center at Xilinx. "Leveraging leading standards based interfaces including NVM Express and PCIe Gen4 allows application developers to rapidly deploy new solutions into the market." "The POWER9 processor from IBM is the perfect CPU for data-intensive workloads like machine learning and now we can feed that processor with data velocity that other systems simply can't match," said Stephen Bates, Chief Technology Officer, Eideticom. "Customers need this to improve the performance of their applications and can utilize this benefit to reduce PCIe lane count and free up slots for other IO devices. No other CPU in production today can offer this kind of performance improvement. Combining this CPU with the next-generation OpenCompute/OpenPOWER server from RackSpace leads to a solution with storage bandwidth that is second to none." "The OpenPOWER Foundation has been aggressively adopting PCIe Gen4 because we see the need for faster storage, network and compute," said Bryan Talik, President of the OpenPOWER Foundation. "OpenPOWER has already demonstrated PCIe Gen4 support with IBM, Mellanox, and Xilinx and we are delighted that Eideticom can now offer fast storage and compute via NVMe over that PCIe Gen4 ecosystem."

https://www.prnewswire.com/news-releases/eideticom-ibm-rackspace-and-xilinx-demonstrate-worlds-first-pcie-gen4-nvm-express-production-ready-system-676532203.html

Sunday, February 11, 2018

NVMe flash hopeful E8 Storage goes software-only route

#Allflash array startup @E8Storage has expanded into reference architecture with the launch of a software-only version. The new product, E8 Storage Software, runs on rack servers from @DellEMC, @Hewlett Packard Enterprise (#HPE) and @Lenovo. Customers can buy SKUs through channel partners, or purchase the software-defined flash directly from E8 Storage as an integrated nonvolatile memory express (#NVMe) appliance.

E8 Storage Software is qualified with Dell EMC PowerEdge R740xd and PowerEdge r64, Hewlett Packard Enterprise ProLiant DL360 and DL380 Gen 10, and Lenovo ThinkSystem SR630. The validated hardware systems need to run Red Hat Enterprise Linux of CentOS 7.3 or higher.

The reference stack includes server chassis with 32 GB to 64 GB of memory, Intel Skylake processors, 24 NVMe U.2 hot-swappable SSDs, two 128 GB RAID 1-enabled M.2 boot SSDs with RAID, and two 100 Gigabit Ethernet (GbE) Mellanox Connect X4-C remote direct memory access network interface cards.

The E8 software-defined flash allows 96 clustered hosts to read and write to shared storage. The vendor’s flagship E8-D24 rack-scale system has dual controllers and scales to 140 TB of effective storage with high-capacity SSDs. E8 Storage also is previewing its E8-X24 block arrays with customers running the IBM Spectrum Scale parallel file system and Oracle Real Application Cluster environments.

Thje recent addition of host-level mirroringenables E8 Storage to market its software-only flash storage on its S10 entry-level appliance to enterprise customers. The S10 has a single-controller and has been used mostly with proofs of concept.

“Customers in database environments want us to fit into their existing disaster recovery environment, rather than running an additional layer.  Larger customers may want to do their own integration. We think smaller customers will still want a (turnkey) appliance,” said Julie Herd, E8 Storage director of technical marketing.

The NVMe standard is based on Peripheral Component Interconnect Express (PCIe) protocol. It is designed to squeeze the most performance from software-defined flash storage. Rather than running traffic through network host bus adapters, an application uses PCIe to talk directly to storage.

NVMe flash storage is maturing to the point that some industry observers predict an uptick in mainstream adoption in 2018. The NVM Express organization, a consortium of industry partners, is expected to help advanced NVMe over Fabrics technologies this year.

E8 Storage and other NVMe flash startupsare jockeying for position, while established vendors Hitachi Vantara, IBM and Pure Storage are bringing systems to market built with custom flash modules.

Herd said the reference architecture stacks will help E8 Storage take on more workloads.  “We are a block system, so this will help us tackle file-based workloads, and also broadens the market for our channel partners.”

http://searchstorage.techtarget.com/blog/Storage-Soup/NVMe-flash-hopeful-E8-Storage-goes-software-only-route

Tuesday, January 23, 2018

NVMe enterprise flash storage, new memory mean change in '18

At the start of 2018, the storage industry is prepared for an influx of ultrafast, low-latency #NVMe -based enterprise #flash storage and even higher-performing new memory technologies. Technology executives at leading vendors and storage analysts predict the industry will begin to shift to NVMe-based PCI Express SSDs in 2018 and eventually to enterprise flash storage arrays that support nonvolatile memory express over Fabrics ( #NVMe-oF ) to extend the performance benefits across a network. On the horizon, they also see new memory technologies, such as #3DXPoint from @Intel and @Micron, @Samsung #ZNAND and other emerging #storage class memory ( #SCM ) options, becoming a major disruptive force in the industry. The lightning-fast memory technologies could spur a new wave of software options and storage rearchitecting to address bottlenecks that have shifted from hardware to legacy software stacks. Below are predictions on enterprise flash storage, NVMe technologies and emerging persistent memory options that SearchStorage collected from industry experts.


Sunday, January 14, 2018

Potential NVMe benefits spark enterprise IT interest

IT pros accustomed to SAS- and SATA-based flash storage performance are eagerly anticipating the latest NVMe technologies. Some are already using nonvolatile memory express for demanding applications, while others wait for vendor support.

NVMe-based solid-state drives (SSDs) can lower latency and boost performance, compared with the standard SAS- and SATA-based SSDs that are the norm in most IT organizations. NVMe over Fabrics (NVMe-oF) can extend the NVMe benefits across a network, between application servers and storage arrays.

Although these are early days for NVMe, some enterprises already made the jump for select application workloads that have especially high I/O requirements. Market statistics indicate those users are in the distinct minority, but adoption should start to increase as storage vendors ramp up support and enterprise NVMe-based PCI Express SSDs drop in price.

NVMe benefits for database response time

Baystate Health took the leap to NVMe-based SSDs last year to try to boost database response time. The healthcare provider, based in Springfield, Mass., uses VMware's vSAN hyper-converged infrastructure on Cisco Unified Computing System server hardware across three data centers in about a 10-mile radius.

"It's not like we want to do cool, fancy things," said Raj Subramanian, director of IT infrastructure and technology at Baystate Health. He said NVMe had become "necessary for the response time."

Subramanian said Baystate's InterSystems Caché database, in particular, requires low response time to prevent corruption and performance issues.

"Overall, NVMe improved the cache performance," Subramanian said, "but it still hasn't been determined if the performance is good enough for Caché's IOPS requirements."

Financial institutions tend to be early adopters of any new technologies that can boost performance. One technical director in infrastructure architecture said his financial services firm used Fusion-io's PCIe-based SSDs before shifting to standard NVMe-based PCIe SSDs as soon as the drives became available.

"It's definitely significantly faster. And the pricing has come down on the NVMe drives, so it's actually pretty competitive, especially in comparison to a SAS flash drive," said the technical director, who asked that neither his name nor company be identified.

The financial services firm bought @Samsung NVMe-based PCIe SSDs for its #Aerospike in-memory #NoSQL database running on @Dell servers, with plans to explore storage-class memory for its trading systems and in-memory databases. The technical director said the firm's other databases use conventional disk- and flash-based SAN storage, and the company's virtual servers connect to @Pure Storage and Dell EMC #XtremIO #allflash arrays.

http://searchstorage.techtarget.com/news/450433021/Potential-NVMe-benefits-spark-enterprise-IT-interest

Thursday, January 11, 2018

Potential NVMe benefits spark enterprise IT interest

Enterprise IT pros who have seen the benefits of conventional flash storage are looking ahead to potential use cases for faster NVMe-based technologies.

IT pros accustomed to SAS- and SATA-based flash storage performance are eagerly anticipating the latest NVMe technologies. Some are already using nonvolatile memory express for demanding applications, while others wait for vendor support.

NVMe-based solid-state drives (SSDs) can lower latency and boost performance, compared with the standard SAS- and SATA-based SSDs that are the norm in most IT organizations. NVMe over Fabrics (NVMe-oF) can extend the NVMe benefits across a network, between application servers and storage arrays.

Although these are early days for NVMe, some enterprises already made the jump for select application workloads that have especially high I/O requirements. Market statistics indicate those users are in the distinct minority, but adoption should start to increase as storage vendors ramp up support and enterprise NVMe-based PCI Express SSDs drop in price.

IT pros accustomed to SAS- and SATA-based flash storage performance are eagerly anticipating the latest NVMe technologies. Some are already using nonvolatile memory express for demanding applications, while others wait for vendor support.

NVMe-based solid-state drives (SSDs) can lower latency and boost performance, compared with the standard SAS- and SATA-based SSDs that are the norm in most IT organizations. NVMe over Fabrics (NVMe-oF) can extend the NVMe benefits across a network, between application servers and storage arrays.

Although these are early days for NVMe, some enterprises already made the jump for select application workloads that have especially high I/O requirements. Market statistics indicate those users are in the distinct minority, but adoption should start to increase as storage vendors ramp up support and enterprise NVMe-based PCI Express SSDs drop in price.

NVMe benefits for database response time

Baystate Health took the leap to NVMe-based SSDs last year to try to boost database response time. The healthcare provider, based in Springfield, Mass., uses VMware's vSAN hyper-converged infrastructure on Cisco Unified Computing System server hardware across three data centers in about a 10-mile radius.

"It's not like we want to do cool, fancy things," said Raj Subramanian, director of IT infrastructure and technology at Baystate Health. He said NVMe had become "necessary for the response time."

Subramanian said Baystate's InterSystems Caché database, in particular, requires low response time to prevent corruption and performance issues.

"Overall, NVMe improved the cache performance," Subramanian said, "but it still hasn't been determined if the performance is good enough for Caché's IOPS requirements."

Financial institutions tend to be early adopters of any new technologies that can boost performance. One technical director in infrastructure architecture said his financial services firm used Fusion-io's PCIe-based SSDs before shifting to standard NVMe-based PCIe SSDs as soon as the drives became available.

"It's definitely significantly faster. And the pricing has come down on the NVMe drives, so it's actually pretty competitive, especially in comparison to a SAS flash drive," said the technical director, who asked that neither his name nor company be identified.

The financial services firm bought @Samsung #NVMe-based PCIe SSDs for its #Aerospike in-memory #NoSQL #database running on @DellEMC #servers, with plans to explore storage-class memory for its trading systems and in-memory databases. The technical director said the firm's other databases use conventional disk- and flash-based SAN storage, and the company's virtual servers connect to @Pure Storage and Dell EMC #XtremIO #allflash arrays.

http://searchstorage.techtarget.com/news/450433021/Potential-NVMe-benefits-spark-enterprise-IT-interest

Tuesday, January 9, 2018

Supermicro Unveils New All-Flash 1U Server that Delivers 300% Better Storage Density with Samsung's Next Generation Small Form Factor (NGSFF) storage

LAS VEGAS, Jan. 8, 2018 /PRNewswire/ --@Super Micro Computer, Inc. (NASDAQ: #SMCI), a global leader in #enterprise #computing, #storage, #networking #solutions and #greencomputing #technology, today unveiled a new #allflash #NVMe™ (Non-Volatile Memory Express) 1U SuperServer that accommodates up to 36 hot-swap @Samsung #NGSFF NVMe SSDs. New Supermicro all-flash 1U Storage Server supports 36 Samsung NGSFF NVMe SSDs. With 36 hot-swap Samsung NGSFF NVMe drives in a 1U system, Supermicro's new solution will provide all-NVMe capacity at petabyte scale.  The server currently supports 288TB per system and will support over half a petabyte per system later this year. NVMe technology was developed to provide the lowest latency and offer faster CPU-to-data storage performance for advanced computing. Optimized for the new Samsung NGSFF SSDs, the new Supermicro system delivers 300% better storage density compared to previously available all-flash storage servers that support standard (U.2) SSDs. "As a market leader in state-of-the-art server architectures, Supermicro recognizes the importance of new protocols such as NVMe and RDMA over Fabric that effectively leverage the outstanding performance characteristics of flash to replace legacy storage form factors and architectures," said Charles Liang, President and CEO of Supermicro. "With our adoption of NVMe and NGSFF, Supermicro continues to demonstrate market leadership by delivering the best performance and density in the market place with these advanced 'future proof' solutions." "We have been working closely with Supermicro to refine our new NGSFF NVMe SSD solution for 288TB to 576TB per system, so data center decision-makers can see that our NGSFF SSD provides the best performance package for all-flash NVMe enterprise storage in 1U systems," said Jim Elliott, Corporate Senior Vice President of Memory Sales and Marketing, Samsung Semiconductor, Inc. "Samsung's technology leadership will continue to drive new applications as we accelerate growth of the NVMe SSD market, where together with Supermicro, we are delivering unparalleled density and performance advantages for IOPS-intensive applications."

https://www.prnewswire.com/news-releases/supermicro-unveils-new-all-flash-1u-server-that-delivers-300-better-storage-density-with-samsungs-next-generation-small-form-factor-ngsff-storage-300578553.html

Wednesday, November 29, 2017

Supermicro Introduces Next-Generation Storage Form Factor with New Intel "Ruler" All-Flash NVMe 1U Server and JBOF

SAN JOSE, Calif., Nov. 28, 2017 /PRNewswire/ -- #SuperMicro Computer, Inc. (NASDAQ: SMCI), a global leader in #enterprisecomputing, #storage, #networking solutions and #greencomputing technology, today announced availability of a new all-flash #NVMe™ (Non-Volatile Memory Express) 1U JBOF (Just a Bunch Of Flash) and 1U #SuperServer with support for 32 hot-swap Intel "Ruler" form factor NVMe SSDs. Supermicro first with 1U all-flash NVMe solution optimized for new Intel “Ruler” form factor SSDs With a total of 32 hot-swap "Ruler" NVMe drives in a 1U system, @Supermicro 's new NVMe solution will provide all-NVMe capacity at petabyte scale in 1U of rack space as the company supports 256TB per system today and will support 512TB and then 1PB per system in the near future. NVMe technology was developed to unleash the best possible latency and provide faster CPU to data storage performance for advanced computing. Optimized for the new Intel® "Ruler" form factor, the new Supermicro system provides a more thermally optimized high-density, high-performance all-flash storage solution compared to previous all-flash storage technologies. "With Supermicro 3.0, our enterprise customers benefit from the industry's broadest selection of first-to-market server and storage solutions with global reach, premium quality, RAS security, rack scale management and global services.  Our new all-flash 32 hot-swap Ruler drives in a high-density 1U system design is the latest example of how Supermicro continues to lead the way for NVMe technology," said Charles Liang, President and CEO of Supermicro.  "With more than triple the all-flash storage density of previous 1U solutions, this Supermicro system will take us to Petabyte scale in a single 1U system in the near future.  This new JBOF supports flexible configurations with up to twelve hosts or head nodes and extremely high data transfer throughput up to 64GB per second." The new 1U all-NVMe Storage Server and JBOF disaggregate storage into shared pools that are rapidly becoming the preferred hardware infrastructure for demanding Big Data analytics applications such as autonomous driving and real-time financial fraud detection.  Up to 12 hosts can be directly connected to the 1U pooled NVMe storage.  Alternatively, for customers who want to deploy an NVMe over Fabric (NVMeoF) solution, hundreds of hosts can be connected to the pooled high-performance NVMe storage over Ethernet. Supermicro 1U all-NVMe Storage Servers and JBOF solutions help maximize high-performance storage resource utilization and reduce the datacenter footprint resulting in lower TCO.      Supermicro's new all-flash 32 drive NVMe 1U system supports not only Intel "ruler" form factor SSDs, but also standard U.2 SSDs to offer customers greater storage flexibility.  This 1U system will support a half petabyte of NVMe storage capacity this year and a full petabyte early next year.  The system comes standard with redundant hot-swap cooling fans and power supplies along with tool-less drive trays for increased serviceability and redundancy. For accessibility, the solution supports remote system on/off and system management as well as remote power cycling for each individual drive.  For more information on this new JBOF, please go to: https://www.supermicro.com/products/system/1U/136/SSG-136R-NR32JBF.cfm This innovative high-end all-flash 1U system is the newest addition to Supermicro's extensive portfolio of industry leading storage servers and JBOD product lines.  With 2U, 3U and 4U offerings that include all-flash NVMe, Simply Double, double-sided and top-loading options with SAS3 RAID or HBA controllers, Supermicro provides the industry's broadest selection of storage products to meet today's stringent customer requirements

https://www.prnewswire.com/news-releases/supermicro-introduces-next-generation-storage-form-factor-with-new-intel-ruler-all-flash-nvme-1u-server-and-jbof-300562353.html

Sunday, November 12, 2017

Huawei Expands NVMe All-Flash Storage Arrays

Storage media has gone through several transformations with the rate of change picking up in recent years. From tape to spinning disks to flash, not only have the style of media changed, each new media type drastically improved upon performance. Flash itself has seen new innovation and performance. Through newly introduced interfaces users have seen performance increase rapidly from SATA to SAS to PCIe and now NVMe. The onus now is on storage array vendors to take advantage of these advances to deliver solutions the enterprise needs as application needs continue to grow.  In terms of performance, NVMe is the current reigning champion in the world of storage. Most major players have either released a version of NVMe media or changed their arrays and servers to be compatible with the interface. The NVMe interface allows the storage media to take advantage of both the low latency of PCIe and the concurrency capabilities of modern CPUs and applications. This results in performance gains that can be further compounded when put in an all-flash platform that is specifically designed to not only handle NVMe, but leverage NVMe SSDs to their fullest. All-flash arrays have been the dominant performance force in the storage array market for a few years now. However, applications such as cloud computing, online payment transactions, social networking and other real-time interactive apps are demanding even higher performance and faster speeds. To meet these demands companies are now looking at the even faster non-volatile memory express (NVMe) capabilities. Huawei is tapping into the technology’s potential to accelerate an increasing proportion of its customers’ workloads. The latest @Huawei flash array is the #OceanStor #Dorado V3 series. This new array takes Huawei’s 12 years of flash research and culminates it into a solution that is ideal for mission-critical workloads. The dual controller array supports up to 25 #NVMe #SSD s (1TB, 2TB, or 4TB with a maximum raw capacity of 100TB per array). Eight OceanStor Dorado5000 V3 arrays can be supported in a single system for a total of 16 controllers and 200 NVMe SSDs, with a total raw capacity of 800TB. The company states that the new OceanStor Dorado5000 V3 with NVMe drives will have 50% higher performance than the previous model leverage SAS SSDs. As is expected leveraging NVMe, organizations will see an noticable uptick in performance combined with a drop in latency. The NVMe OceanStor Dorado5000 V3 pshes performance as high as 4 million IOPS, storage latency as low as 0.5ms, and up to 6 nines of availability. The new array also falls under Huawei’s data reduction guarantee of 3:1, which can lower TCO up to 50% compared to less efficient storage. With Huawei, customers get a vertically integrated solution, which is somewhat unique in the array space. The company does all of the work in house including the SSDs, NAND, and the controllers. Huawei has also used the NVMe over Fabric (NVMeoF) 1.0 protocol, released in June of 2016, to more effectively take advantage of the unique benefits of NVMe. They have separated the control and I/O paths to reap the benefits while avoiding pitfalls like over-taxing the CPUs. Like the rest of the OceanStor Dorado V3 family, the new array will come with Huawei's proprietary FlashLink technology that includes on-board I/O priority scheduling adjustment, large block sequential write, global garbage collection, and other utilities to achieve predictable 0.5ms latency. When paired with Huawei HyperMetro gateway-free array active-active solution, the array can hit six-nines (99.9999%) reliability. The new array also comes with inline deduplication and compression technology to help companies get the most bang out of their storage buck. Another feature of the new OceanStor Dorado5000 V3 is the addition of Huawei’s new Dedicated Enterprise Storage Service (DESS). A big focus of the company’s over the past few years has been the cloud. Huawei is looking to help organizations move to the cloud or begin there. DESS acts as an enterprise-grade storage-as-a-service (STaaS) for painless migration of mission critical services to the cloud. Being in the cloud allows these services to performance needed as well as the latest technology as it becomes available. DESS can also provide cloud-based data backup, hot/cold data tiering, and a long list of other utilities.

http://www.storagereview.com/huawei_expands_nvme_allflash_storage_arrays

Saturday, October 14, 2017

Get ready for Gen 6 Fibre Channel technology

What's driving 32-gig Fibre Channel adoption? #HyperConvergence and the #cloud favor Ethernet over Fibre Channel ( #EoFC ) technology. And Ethernet is better for #unstructureddata storage, which is more popular than the #structureddata that needs FC #SAN s. So why does anyone care about 32-gig FC? Flash. #Allflash storage SANs could keep Fibre Channel in the picture, especially while the technology transitions from 16 Gbps to 32 Gbps equipment. As performance bottlenecks move from arrays to the storage network, FC will stay relevant. Because solid-state storage has greater throughput and lower latency than hard disks, storage networks require more bandwidth to meet performance peaks. Gen 5 Fibre Channel will eventually be too slow for faster SSDs, such as flash with nonvolatile memory express ( #NVMe ) and #3DXPoint. Experts agree that as drive performance improves, networks must receive upgrades to 32-gig FC and 100-gig Ethernet. That said, 32-gig adoption has been slow. Vendors such as @Brocade, @Cisco and @QLogic offer switches and adapters to support #32gig FC, but it's possible that organizations are waiting for the price of #allflash arrays to come down. Companies that get on the 32-gig bandwagon early can still use their 16-gig equipment, but organizations looking on could be holding out until all the kinks are worked out. What will truly start to drive 32-gig FC adoption will be array vendors. Once all-flash arrays embrace Gen 6 Fibre Channel, organizations upgrading to all-flash will too. In the meantime, learn everything you need to know about 32-gig Fibre Channel, NVMe and more in our three-part guide to Gen 6 Fibre Channel technology.
http://searchstorage.techtarget.com/ehandbook/Get-ready-for-Gen-6-Fibre-Channel-technology

Thursday, October 12, 2017

Samsung: NVMe benefits are available now, within limits

#NVMe can bring significant benefits in performance and power consumption today, but the new #flashstorage protocol is set for real take off in 2018.
Those are the views of the solid state drive product marketing executive for @SamsungSemiconductor Europe, Kristian Vättö, who told Computer Weekly NVMe will take off as a storage medium in enterprise arrays with the arrival of dual port drives, next generation central processing units (CPUs) and NVMe-over-fabrics (#NVMf ).

The key problem highlighted by Vättö is that currently NAS and SAN have to switch between different protocols when they access storage, transitioning between small computer system interface (SCSI) and Ethernet and Fibre Channel.

“Because they are engaged in protocol switching, latency is added,” said Vättö. “That wasn’t a problem with hard disk drives when their latency was higher than that of the network.”

NVMe is an emerging flash storage protocol that can potentially speed access times by orders of magnitude compared to the SCSI and SATA protocols designed for spinning disk, but still in use for many flash drives.

But, hardware makers need to eliminate SCSI from the input/output (I/O) path and a bottleneck in storage processing to allow NVMe off the leash.

According to Vättö, NVMe’s most successful early adopters are hyperscaleusers. These are the likes of Google, Facebook etc that design their own hardware and software for use in their datacentres and don’t use proprietary storage platforms.

http://www.computerweekly.com/news/450428025/Samsung-NVMe-benefits-are-available-now-within-limits

Wednesday, September 20, 2017

NetApp Introduces NVMe Over InfiniBand For Big Data Analytics

Today #NetApp introduced new offerings, including both new solutions and software, to its #BigData analytics portfolio. Amongst these new offerings is the NetApp #ConvergedInfrastructure Solution for #DataAnalytics, the first available #NVMe-enabled solution in this particular market. The new offerings bring cost-effectiveness and ultrafast performance to #BigData analytics.  Big Data, and the insights that can be gained from it, has been a central issue for some time now. However, there doesn’t seem to be too many solutions out there that can analyze Big Data in a time-effective manner to gain insights that can leveraged before the next big wave of data comes crushing in. Meaningful analytics should be the goal, and that is exactly what NetApp is promising with its new offerings. The new solutions and software introduced include: NetApp EF570 all flash system – is the storage industry’s price/performance leader as demonstrated by SPC-1 and SPC-2 benchmark tests. The 4th generation all-flash array designed specifically for performance-intensive workloads such as big data analytics, technical computing and video surveillance, now delivers up to 21GBps of bandwidth, 1M sustained IOPS, sub-100 microsecond latency, and support for 100Gb NVMe over InfiniBand to increase application responsiveness and accelerate analytics – all in a compact, 2U package. The system supports 367TB of raw flash capacity in a modular 2U building block that scales to 1.8PB. NetApp E5700 Series hybrid-flash system – accelerates performance and increases density with a hybrid flash system that is perfectly suited for modern enterprise applications. It also supports 100Gb NVMe over InfiniBand. SANtricity Cloud Connector – enables Data Fabric capabilities by providing cost-effective backup and recovery to the cloud from E/EF-Series systems. NetApp SANtricitry software – SANtricity 11.4 now comes with security enhancements such as role-based access control and audit log, Active Directory support, an external key manager, and ships with on-box SANtricity System Manager. NetApp Converged Infrastructure Solution for Data Analytics – brings together the latest storage, networking, and server technologies in a pretested, prevalidated reference architecture to help simplify deployment of data analytics environments. The new all-flash EF570 is aimed at the midrange. It now comes with 2x the throughput of the previous version. This new performance can, for example, reduce video processing by as much as 95%, reduce data analysis from weeks to hours, and quadruple response times for financial transactions. The hybrid E5700 is designed with NetApp SANtricity operating system adaptive caching algorithms for a large range of application workloads. According to NetApp, the E5700 combines extremely high IOPS with sub-100 microsecond response times to increase application responsiveness and accelerate analytics, all while taking up either 2U or 4U of space. Through integration with NetApp Data Fabric, customers will see seamless backup and recovery to the hybrid cloud. Both the all-flash EF570 and hybrid E5700 support multiple high-speed host interfaces including 32Gb Fibre Channel, 25Gb iSCSI, 100Gb InfiniBand, 12Gb SAS, and 100Gb NVMe over InfiniBand, making the EF570 and E5700 Series one of the first 2U arrays with support for NVMe. Both series leverage SANtricity software and offer extreme performance, unmatched value, and proven simplicity making them ideal for backup and recovery, technical computing, video surveillance, and big data analytics. Availability and Pricing The EF570 starts at $25K with 12 800GB SSDs and is expected to be available in October.

http://www.storagereview.com/netapp_introduces_nvme_over_infiniband_for_big_data_analytics

Sunday, September 10, 2017

Learn more about the all-flash storage vendors that appeared in the latest Gartner Magic Quadrant report.

Gartner recently released a Magic Quadrant (MQ) for all-flash storage arrays. Gartner analyst Valdis Filks said the overall solid-state array (SSA) market grew by 72 percent and generated revenues of $4.6 billion in 2016. This growth is being fueled in part by the purchase price dropping 40 percent or more in the last 18 months. “Most of the traditional incumbent and system vendors have transitioned their overall storage array businesses to SSAs, where more than 50 percent of new storage array sales are SSAs,” said Filks. “Solid-state arrays have moved to highly integrated application development platforms with agility features such as many different types of data copies, backups, migrations and services for applications within containers.”  Here's in-depth look at the top ten vendors from the MQ. Pure Storage Pure Storage achieved pride of place among vendors in the Leaders quadrant. This is due to what the analyst firm considers the completeness of its vision as well as its ability to execute. Part of the reason for this is that Pure Storage didn’t exist until 2007. It has only ever built and sold all-flash arrays. Pure has two primary product families: FlashArray and FlashBlade. Within the FlashArray family are FlashArray//M, optimized for accelerating any business application and structured workloads, and FlashArray//X. “Our FlashArray family of all-flash block storage arrays are deployed all the way from tier 1, mission-critical environments running business and transactional workloads such as Oracle, SQL and SAP, through to deeply consolidated and automated private and hybrid cloud environments,” said Jason Nadeau, senior director of products, Pure Storage. “In addition, FlashArray//M serves as the storage component for FlashStack, our converged offering with Cisco.” FlashArray//X is the company’s newest offering and is an all-NVMe array. Existing FlashArray//M systems, meantime, are engineered to be upgradeable to NVMe. The company expects NVMe to follow a similar trajectory to flash itself — an expensive, niche media to start, which is then democratized and becomes industry standard. Overall, FlashArray offers capacity from 5 TB to 512 TB raw (from 15 TB to 1.5 PB effective), a data reduction average of 11:1 (including thin) or 5:1 (not including thin). and pricing starting at $50,000. FlashBlade, optimized for unstructured data, is finding traction in life sciences/genomics, healthcare, self-driving car simulations, rockets and supercomputing. It tends to be sold as much into research and development teams as traditional IT functions, with an emphasis on emerging workloads like artificial intelligence, analytics and machine learning. It offers capacity from 56 TB to 3.9 PB raw (from 98 TB to 8 PB effective), a data reduction average of 3:1, bandwidth from 7 GB/s up to 75 GB/s, and pricing starting at $150,000. Dell EMC #DellEMC is another vendor includes in Gartner’s Leaders quadrant. Its flash storage options include deduplication, compression, thin provisioning, snapshots, encryption, replication and other security features. “With all-flash, you get the speed, agility and scale you need to react to changing markets and customer demands,’ said Jeff Boudreau, president, Dell EMC Storage Division. The Dell EMC all-flash portfolio was recently refreshed with midrange (Dell EMC #Unity, #SCSeries), high-end ( #XtremIO and #VMAX ) and #softwaredefined ( #ScaleIO ) products, as well as unstructured data storage ( #Isilon ). The new VMAX 950F is said to be 68 percent faster and offer 30 percent better response times (6.7 million IOPS, 350 microseconds response times for OLTP) than the previous generation (VMAX 850F). It is best for the consolidation of mission-critical applications, such as core banking, credit card processing, electronic billing or hospital record systems. The XtremIO X2 is designed for virtual desktop infrastructure (VDI) and large-scale snapshot use cases. It is said to have three times higher capacity, 25 percent better storage efficiency on average and 80 percent better response times than first-gen XtremIO. The X2’s rack density has been increased by four times from the previous generation, providing up to 5.5 PB effective capacity and a capacity density of over 100 TB effective per rack unit. Further, Dell EMC Unity All-Flash storage models feature up to four times larger file system capacity (up to 500 TB), eight times the density and sub-10-minute deployment compared to the previous generation. “Dell EMC flash storage arrays have been repositioned to move the block-only XtremIO family into more specific niche market segments,” said Filks. “The unified (block and file) VMAX All-Flash and Unity All-Flash families are being strategically positioned for diverse high- to midrange applications, and the Isilon file and object systems now have a full-flash model.” HPE Hewlett Packard Enterprise (HPE) is yet another Leader, according to Gartner. It has a couple of all-flash product lines. Nimble Predictive AF units are at the high end and can be deployed in multi-petabyte environments. Gartner said that user satisfaction from this technology (acquired from Nimble) is high. In addition, InfoSight Predictive Analytics offers remote monitoring and analytics to identify issues within and outside of the storage infrastructure. HPE 3PAR StoreServ arrays are aimed at the lower end of the market. Although larger 3PAR units have good capacity and IOPS capabilities, the smallest 3PAR configuration is best suited for environments with modest capacity and IOPS requirements. Both product lines feature deduplication and compression, as well as scale-up or scale-out options. They can take advantage of NVMe Peripheral Component Interface Express (PCIe) SSDs and storage-class memory without a major upgrade. And with HPE now aggressively rolling out Nimble products, they should gain more traction by taking advantage of a strong HPE channel partner ecosystem. NetApp NetApp is another company that Gartner classifies as a Leader. Its All-Flash FAS A-Series (AFF) is the high-end line. It includes ONTAP data management and comes with 15 TB SSDs and up to 7 million IOPS per cluster with sub-millisecond latency, plus 40 Gigabit Ethernet (40GbE) and 32Gb Fibre Channel connectivity. Inline deduplication and data compaction are also part of the package. In addition, SolidFire all-flash arrays are said to provide guaranteed performance to thousands of apps on one shared storage system. Users can mix and match nodes to right-size performance and scalability. Finally, NetApp EF-Series all-flash arrays run on the SANtricity software platform, which is aimed at environments with extreme demands for latency-sensitive workloads. EF-Series products include automated path failover, online administration, data protection, proactive monitoring and repair, nondisruptive upgrades and diagnostic capabilities. IBM Storage IBM has a large collection of solid state arrays, which helped propel the company into Gartner’s Leaders quadrant. IBM SSAs include the A9000, DS8000 series, Storwize series and the V9000. The IBM FlashSystem A9000 series leverages XIV storage system software technology, as well as deduplication and compression. FlashSystem products use IBM's flash module technology to optimize flash drive components for performance and reliability. Software-defined storage features from IBM Spectrum Storage offer additional functions, such as hybrid cloud support. “IBM has a broad portfolio spanning high-performance customer needs, including mainframe environments as well as entry-level and midrange products,” said Filks. “IBM Spectrum Storage software suite allows customers the flexibility to manage IBM and/or external storage solutions, and hybrid cloud support.” IBM’s family of all flash arrays are available to enterprises and small companies. They range from ultra-low latency (FlashSystem 900, FlashSystem V9000, FlashSystem A9000, FlashSystem A9000R, Storwize V7000) to more cost-effective models such as the Storwize V5030F and Storwize V7000F. “With the maturity level that flash has reached in the past 24 months, flash can be deployed in any size company for their primary storage needs,” said Eric Herzog, vice president of worldwide storage channels, IBM. “Spectrum Virtualize software allows a client to improve hybrid cloud and storage deployment flexibility, enjoy consistent management for on-premises and cloud storage regardless of storage type, replicate from on-premises storage to cloud storage.” IBM Spectrum Virtualize software supports almost 400 different storage systems. Further features include storage pooling, automated allocation with thin provisioning, automated tiering, software encryption, local and remote replication, support for moving snapshots from existing storage to public cloud providers, data mobility between data centers and support for virtual environments. Hitachi Data Systems Hitachi Data Systems (HDS) is another vendor that sits in the Leaders quadrant. The HDS Virtual Storage Platform (VSP) includes the VSP F400, F600, F800 and the newest F1500. Late last year, it also brought in 14 TB flash modules. This doubled the amount of raw storage capacity. At the same time, it dropped the Hitachi Flash Storage (HFS) A-series line. According to Gartner, HDS has high satisfaction, support and quality ratings. The VSP line comes with array software included. But its advanced software package, including Hitachi Data Instance Director (HDID) application-aware copy management, is sold at a premium. Standard features include block and file protocols, file deduplication and storage virtualization. “Unlike most storage vendors that need a storage gateway, the HDS VSP F series natively integrates with Amazon Web Services (AWS) S3, Microsoft Azure and private cloud solutions on Hitachi Content Platform, and enables automatic migration of file data to these platforms,” said Filks. Kaminario The Kaminario K2 Gen6 is a scale-up and scale-out all-flash storage system that is said to support flexibility, performance and cost efficiency. Performance scales to over 1.5 million IOPS and more than 25 GB/sec throughput at low latency. The K2 has adaptive block size technology that enables support for mixed workload environments (such as analytics and transactional processing running at the same time). It also supports 87.5 percent utilization RAID and has inline global and selective deduplication, as well as inline byte-aligned compression. It is run by the Kaminario VisionOS distributed storage software stack. This includes enterprise data services for tier 1 storage, such as replication, snapshots, management and non-disruptive operations. Kaminario recently announced Kaminario K2.N and Flex. K2.N is based on VisionOS with an NVMeF backend and optional front end. This makes it possible to disaggregate capacity and compute to allow more flexibility in how storage is provisioned. Kaminario Flex is a storage management and orchestration layer managing K2.N resource pools to compose virtual private arrays. It is driven by the Kaminario analytics platform Clarity. K2.N and Flex will be available in the first and second half of 2018 respectively. Kaminario believes that the next-generation storage paradigm will be ruled by having the right software storage stack to meet the needs of next-generation applications. That means supporting extreme flexibility and performance at scale while maintaining cost efficiency and delivering a full set of data services. Any such platform must continually adopt the latest hardware and deliver it in a seamless integrated way to the customers. It touts its VisionOS as a way to achieve this. “Kaminario is focused on the SaaS, online, business and private cloud market,” said CTO Eyal David. “These customers have the need for the scalability, performance and agility that our platform provides.” Gartner classified Kaminario as a Leader. It gave the company credit for adding an NVMe/PCIe-connected compression offload engine, which is ready for more NVMe and NVMe fabric features as they are developed. In particular, Gartner noted that Kaminario has technological NVMe leadership and execution. Tegile Gartner classifies Tegile as a Visionary in the bottom right quadrant. It earned that label due to the fact that many of the guarantees, purchase options, features and programs that Tegile offers cannot be matched by competitors. “The interoperability, flexibility of options and features of the Tegile arrays are probably the most extensive in the market,” said Filks. “The arrays offer unified protocol support, very detailed instrumentation and configuration flexibility.” The Tegile IntelliFlash HD is aimed at medium and large enterprises. It is said to deliver up to five million IOPS and 10 PB of effective capacity in a rack. It can be used to consolidate mixed workloads into a single platform. Additionally, the company just updated its all-flash product line for the NVMe market, Gartner expects to be worth more than $50 billion by 2020 due to the demand for speed and performance. The Tegile N Series all-flash storage arrays incorporate solid state media with NVMe technology (24 HGST Ultrastar NVMe SSDs in a 2U footprint). This can offer up to 3 million IOPS with consistent 200 microsecond latency. “With the introduction of NVMe to the market, the next major battle in the enterprise storage market will be fought in the realm of persistent memory storage,” said Rohit Kshetrapal, CEO of Tegile. Tintri Tintri didn’t make it into the Leaders quadrant, but it did meet it into the Visionaries section. Filks noted that Tintri’s focus is on server-virtualized environments. With a new IPO behind it, Tintri is expanding. On the product front, it has announced plans for a product refresh and larger arrays later in the year. Currently, there are three T5000 series models available: T5040, T5060 and T5080. They scale from 6 TB at the low end to 92 TB of raw capacity at the top. Expect those numbers to shift significantly when the new models come out. The company claims up to 30:1 data reduction numbers. The Tintri SSAs also have hypervisor integration, simple administration and provisioning, as well as deep integration with VMware, Microsoft Hyper-V and OpenStack. On the downside, there is no support for block storage or physical hosts. Fujitsu Gartner classes Fujitsu as a Challenger when it comes to all-flash storage. The Fujitsu ETERNUS AF storage product line is said to be affordable and easily configurable. It has an automated quality of service feature to ensure each application gets the performance it needs. Deduplication and compression are included, as well as disaster recovery (DR), mirroring and automated transparent failover. The ETERNUS AF line consists of two models: the entry-level ETERNUS AF 250 (>700 TB SSD capacity) and the midrange ETERNUS AF650 (up to 3 PB of SSD capacity). “The key differentiators of Fujitsu ETERNUS AF systems are response time, flexible deduplication and compression, automated quality of service management and transparent failover,” said Frank Reichart, head of storage marketing at Fujitsu. “If an ETERNUS AF system goes down, an ETERNUS AF or DX system can take over within five seconds at another location.” Gartner drew attention to the fact that the Eternus AF250 (with 737 TB of raw capacity) and Eternus AF650 (with 2.9 PB raw capacity) are backward compatible with earlier Fujitsu flash arrays, include good migration services from Fujitsu HDD arrays to SSAs, and have beneficial swap-out financial terms for Fujitsu-sold Violin Memory arrays.


Wednesday, September 6, 2017

Huawei developing NMVe over IP SSD

Analysis #Huawei is developing an #NVMeoverIPSSD with an #ondrive #objectstorage scheme meaning radically faster object storage and a re-evaluation of what object storage's very purpose. At the Huawei Connect 2017 event in Shanghai, Guangbin Meng, Storage Product Line President for Huawei, told El Reg Huawei is developing an NVMe over IP SSD to overcome in-system scaling limits. Such an SSD could enable an all-flash array to scale up to tens of thousands of nodes. Each drive would have its own IP address. This reminds El Registero of Seagate’s Kinectic disk drive idea, in which each disk drive has its own Ethernet address and implements an object storage scheme with Get and Put data access operators. The Kinetic disk concept seems to have run its course inside Seagate and facing oblivion. However startup OpenIO is persevering with object storage disks. Igneous has similar ideas with Ethernet-Accessed, ARM CPU-driven disks. Cameron Bahar, VP and CTO for Enterprise Storage at Huawei’s Santa Clara facility, says Huawei actually had the Kinetic drive idea before Seagate. He explains that Garth Gibson, the man who developed RAID proposed a NAS-attached storage disk in a pre-2000 paper. He then went to Panasas as CTO and tried to develop the idea but it didn’t get any traction. In 2012 Huawei built an object store in its Santa Clara facility using key:value disks. Each disk had a daughter card attached to it with an ARM CPU, DRAM and Linux. The device was even patented. However a Huawei employee went to Seagate and persuaded Seagate to develop its Kinetic disk. The drives there already had ARM CPUs inside them looking after the internal disk drive functions, Bahar said the extra cost for a Kinetic drive, the Kinetic tax, was US$20. That's quite high, especially for those who buy disks by the thousand as you would when building the kind of scale-out rigs at which object storage excels. Back to Huawei and its NVMe over IP concept for flash drives; SSDs. Huawei had the idea of combining this with the on-drive object store and creating an object storage system composed from individual flash drive nodes. The object storage software is Huawei’s own and involves a key:value store on each drive. Bahar talked about two kinds of object store; cheap and deep S3-style which is the traditional idea, and fast object stores based on key:value store flash drive nodes. What happens if we have very fast object stores? Let’s put them in an array and add a NAS gateway. We then have a flash-based and vastly-scalable filer. There could be different types of flash; faster and slower drives for example. How enterprises or cloud service providers could use such an array is a fascinating question. We might imagine video production, security surveillance systems, big data analytics and AI/machine learning applications might have an interest in it. It’s only Huawei at present, with a developing and dramatically faster object storage system based on NVME over IP-accessed flash drive nodes. It’s a promising concept that could find traction. Let’s see where it goes

https://www.theregister.co.uk/2017/09/06/huawei_developing_nmve_over_ip_ssd/

Thursday, August 31, 2017

Scale unveils ludicrously fast HCIA with 20μs latencies

#ScaleComputing has run lab tests showing #hyperconverged virtual machine IO performance using #NVM express ( #NVMe ), achieving mean #IOlatencies as low as #20μs delivered to a guest virtual machine. Scale sells #HC1000, #HC2000, #HC4000 and #HC5000 series #hyperconvergedinfrastructure appliances ( #HCIA ) with a #HyperCore operating system and #SCRIBE software storage engine. Scale says guest virtual machines may communicate directly with SCRIBE (using standard Virtio compatible drivers) without involving a hypervisor in any way. This reduces CPU bottlenecks, context switches and system calls, all of which would add latency to each IO. The company has been checking out the use of NVMe flash and 3D XPoint drives and announced the lab test results at VMworld, which is a bit cheeky since it uses the KVM hypervisor. The results are good enough for it to be introducing HyperCore-Direct products, using NVMe drives, in the next few months. Testing of HyperCore-Direct has shown that apps running with VMs can expect near bare-metal NVMe IO performance from mirrored, highly-available, software-defined storage. An example 4-node HyperCore-Direct configuration with all-flash media is able to provide mean IO latencies of 150μs on a mixed (90 per cent read / 10 per cent write) random workload to 24 virtual machines (at a total 240,000 IOPS). With faster Opane media, the same workload achieves mean latencies of 65 microseconds. The 4-node all-flash system has also demonstrated over 2.6m IOPS with under 290μs mean latency, as delivered to a virtualised benchmark application. Windows 7 running IOMeter on a HyperCore-Direct system is able to obtain average latencies of 101μs for a 4k random read workload with a queue depth of 1. Frank Ober, an Intel Optane technology solution architect, said: “Scale Computing’s solution using Intel Optane SSD P4800X is approaching the 10µs or lower bare metal latency for classic virtualisation.” A canned quote from Howard Marks, chief scientist at DeepStorage, LLC, adds: “Scale Computing’s new HyperCore-Direct solution brings the ultra-low latency performance of NVMe to the masses. Less than a year ago, users requiring the sub-500µs latency and low overhead NVMe promised were limited to NVMe SSDs in servers or leading-edge Tier-0 systems that had no data services. Scale’s now delivering that performance on a hyperconverged platform with snapshots, and the other services enterprise applications require.” This is possibly the fastest HCIA system seen in the industry. More info on the test results here, where you can check out a HyperCore-Direct white paper.

https://www.theregister.co.uk/2017/08/30/scales_ludicrous_mode_hcia_with_20_microsecond_responsiveness/

Tuesday, August 29, 2017

Supermicro Accelerates Software-Defined Storage with VMware vSAN Hyper-Converged Solution

LAS VEGAS, Aug. 28, 2017 /PRNewswire/ -- #SuperMicro Computer, Inc. (NASDAQ: SMCI), a global leader in computing, storage, and networking technology and green computing, today introduced its latest fully configured, ready to deploy #softwaredefinedstorage ( #SDS ) solution at #VMworld 2017 in Las Vegas, booth 1113, Mandalay Bay Hotel & Convention Center. "Our new All-Flash #NVMe X11 #BigTwin™ solution is fully ready to deploy, and has undergone stringent validation testing and certification by Intel, VMware and Supermicro," said Charles Liang, President and CEO of Supermicro. "This Intel Select All-Flash vSAN hyper-converged solution accelerates software-defined storage, optimizes data center infrastructure, and ultimately fast-tracks the process of selecting and deploying hardware and software." This Supermicro hyper-converged solution is the result of collaborative efforts with Intel to optimize the performance for workload specific use cases. As part of the Intel Select Solutions program, which is a conglomerate of workload-optimized data center solutions designed to simplify and accelerate the process of selecting and deploying hardware and software, these Supermicro X11 solutions run on Intel® Xeon® Scalable processors. By leveraging Intel Xeon Scalable processors, this BigTwin SDS solution has higher core counts designed to support more virtual machines (VM) per node and system. With up to 3TB memory per node, All-Flash NVMe X11 BigTwin technology is ideal for in-memory database applications. Each node supports six hot-swap NVMe drives to fully maximize performance. Featuring Supermicro's flexible, cost-optimized SIOM networking module, the All-Flash NVMe X11 BigTwin supports a variety of 10G, 25G, 50G and 100G network interface options. The solution's multi-node architecture maximizes power efficiency through resource sharing for greener computing and saves customers on both CAPEX and OPEX to deliver the best TCO. Using #IntelOptane™ #NVMe #SSD s for caching helps this high-performance #SDS solution accelerate storage performance up to 3X with endurance up to 30 drive writes per day (DWPD). These #Supermicro #VMWare solutions based on the #BigTwin offer #vSAN #dataprotection for mission-critical applications.
http://www.prnewswire.com/news-releases/supermicro-accelerates-software-defined-storage-with-vmware-vsan-hyper-converged-solution-300509913.html

Sunday, August 27, 2017

Hedvig Announces Its Distributed Storage Platform 3.0

Today #Hedvig announced the latest release of its software-defined storage ( #SDS ) platform, Hedvig #DistributedStoragePlatform version 3.0. The latest update contains innovations including end-to-end integrations, advanced security capabilities, and enhancements to the platfrom’s comprehensive suite of caching technologies. According to Hevig, these new capabilities will enable customers to store and protect virtualized, containerized, and backup workloads from a single platform.  More and more companies are looking to both SDS and multi-cloud platforms to store both their primary and secondary data. However, these new technologies requires new certifications, plugins, encryption, auditing, multitenancy, and flash-caching capabilities all of which can be found in the new Hedvig Distributed Storage Platform 3.0. The newly updated platform does this through new #CloudScale Plugins, #Hedvig #Encrypt360, and Hedvig #FlashFabric. The company has added new plugins to its existing #Docker and #OpenStack #CloudScale Plugins. CloudScale Plugins provide pretested, validated options that aim to make #SDS easier to use and operate. CloudScale Plugins now include: #Veritas #OpenStorage Technology ( #OST ) plugin. Hedvig’s new OST plugin ensures existing Veritas customers can seamlessly connect #NetBackup to Hedvig as a deduplicating backup target. Enhanced #VMware plugin. The existing Hedvig VMware #vSphereWebClient plugin has achieved VMware-ready storage certification and adds new security, backup, and data protection capabilities. #RedHat Certified #container. A Hedvig Storage Proxy container is now Red Hat certified for customer production use and published in the Red Hat Container Catalog. This certified container supports Red Hat Enterprise Linux (RHEL) and OpenShift featuring docker and Kubernetes. Cyberattacks, including mal- and ransomware, seem to be in the news every other week. To this end, Hedvig is introducing its new Encrypt360 that will deliver native, in-software approach to protecting data throughout its entire lifecycle. Encrypt360 can encrypt data that’s in-use, in-flight, and at-rest, encryption starts at the host level and carries all the way through the backend, distributed cluster nodes. This new security feature supports AWS and allows for a 256-bit AES encryption policy on a per-volume basis. Encrypt360 can be combined with other new features announced today, advanced auditing and multitenancy access control mechanisms, allowing customers to meet compliance and regulations while staying secure in hybrid and multi-cloud architectures. On the flash-caching front, Hedvig is introducing a comprehensive suite of flash caching technologies that optimize performance in Hedvig clusters, FlashFabric. The company has added improvements such as advanced auto-tiering and read cache capabilities. The improvements also allow customers to take advantage of NVMe, 3D Xpoint and other flash innovations that may be announced soon without having to add complex, proprietary hardware.

http://www.storagereview.com/hedvig_announces_its_distributed_storage_platform_30

Thursday, August 24, 2017

Supermicro Delivers Groundbreaking 18 Million IOPS of Storage Performance in New 2U Ultra Server

SAN JOSE, Calif., Aug. 23, 2017 /PRNewswire/ -- #SuperMicro Computer, Inc. (NASDAQ: #SMCI ), a global leader in #compute, #storage, and #networking technology and #greencomputing, today announced the availability of new #allflash #NVMe™ (Non-Volatile Memory Express) 1U and 2U Ultra SuperServers. With support for 20 hot-swap NVMe SSDs, Supermicro's new all-flash 2U Ultra server delivers a groundbreaking 18 million IOPS of storage performance.

"To achieve the lowest possible latency, Supermicro's new all-flash 1U and 2U Ultra servers are designed to support 20 directly attached hot-swap NVMe SSDs," said Charles Liang, President and CEO of Supermicro. "These new X11 servers feature a non-blocking design, allocating 80 PCI-E lanes to the 20 NVMe SSDs for uncompromised Gen 3 PCI-E x4 direct connections that achieve maximum storage performance."

With double the number of directly attached hot-swap NVMe SSDs compared to the previous generation, Supermicro's new all-flash 1U Ultra server supports 20 hot-swap NVMe SSDs.  In addition, the performance-optimized new all-flash 2U Ultra unleashes up to 18 million IOPS of storage performance.

Supermicro X11 Ultra servers fully support the highest end Intel® Xeon® Scalable processors up to 205 watts and 24 DIMMs making these servers an excellent choice for high-performance analytics and in-memory application acceleration. The system architecture is balanced to make optimal use of system resources with each processor supporting 10 NVMe drives and dual 25G ports or a 100G port.

For customers looking for even greater increases in the NVMe storage capacity per rack unit, Supermicro's new 32 hot-swap NVMe SSDs in 1U JBOF (just a bunch of flash) provides maximum high-performance storage density.

With over 100 NVMe based platforms in its X11 server and storage portfolio, Supermicro continues to extend its technology innovation leadership position in NVMe server and storage solutions. 

http://www.prnewswire.com/news-releases/supermicro-delivers-groundbreaking-18-million-iops-of-storage-performance-in-new-2u-ultra-server-300508258.html